使用 Catalog API 对数十亿商品进行聚类以支持智能体商务

Picture two merchants that sell the same protein powder. One creates a single product listing with flavor variants. The other creates a separate listing for every flavor. Both are right. And on their own storefronts, neither structure causes any problems.

想象两个销售相同蛋白粉的商家。一个创建了一个包含口味变体的单一产品列表。另一个则为每种口味创建单独的列表。两者都是正确的。而且在他们各自的店铺中，这两种结构都不会引起任何问题。

But Shopify hosts billions of product listings across millions of stores, and there's no shared schema between them. When an AI shopping agent needs to find the best protein powder across the whole dataset, it has to understand that one merchant's single listing and another merchant's twelve listings describe the same product line… without anyone telling it so.

但是，Shopify在数百万个店铺中托管了数十亿个产品列表，并且它们之间没有共享的Schema。当AI购物代理需要在整个数据集中寻找最好的蛋白粉时，它必须理解一个商家的单个列表和另一个商家的十二个列表描述的是同一个产品系列……而无需任何人告诉它。

This is a tale of teaching machines to read product data the way a human shopper would, at scale.

这是一个关于如何大规模地教导机器像人类购物者一样阅读产品数据的故事。

Enter the Shopify Catalog: a unified intelligence layer that standardizes product data across the platform and makes it available to developers and AI agents via the Catalog API. (Take a look at all of Catalog's new features coming out of our Spring 2026 Edition.)

引入 Shopify Catalog：一个统一的智能层，用于标准化整个平台的产品数据，并通过 Catalog API 将其提供给开发者和 AI 代理。（来看看我们的 Spring 2026 Edition 中推出的所有 Catalog 新功能吧。）

At the heart of Catalog is product clustering. To understand it, it helps to know how Shopify's data model works. Merchants organize their offerings into products (e.g., "Classic Denim Flare Jean"), each of which can have multiple variants (e.g., size 28 in light wash, size 30 in dark wash). Different merchants may structure the same real-world item very differently. One might create a single product with all size and color variants, another might create separate products per color.

Catalog的核心是产品聚类。要理解它，了解Shopify的数据模型是如何工作的会有所帮助。商家将其提供的商品组织成产品（例如，“经典牛仔微喇牛仔裤”），每个产品可以有多个变体（例如，浅色水洗28码，深色水洗30码）。不同的商家对同一现实世界商品的结构化方式可能大相径庭。一个商家可能会创建一个包含所有尺码和颜色变体的单一产品，而另一个商家可能会按颜色创建独立的产品。

Clustering allows us...